Doubt about data augmentation

[BrunoKrinski]

Hi!

I will perform some experiments with data augmentation and I need to train the network without any data augmentation as baseline. I would like to confirm with removing the following lines of semantic.yml file will perform the train without any data augmentation:

    - &hard_transforms
      transform: A.Compose
      transforms:
        - transform: A.ShiftScaleRotate
          shift_limit: 0.1
          scale_limit: 0.1
          rotate_limit: 15
          border_mode: 2  # cv2.BORDER_REFLECT
        - transform: A.OneOf
          transforms:
            - transform: A.HueSaturationValue
            - transform: A.ToGray
            - transform: A.RGBShift
            - transform: A.ChannelShuffle
        - transform: A.RandomBrightnessContrast
          brightness_limit: 0.5
          contrast_limit: 0.5
        - transform: A.RandomGamma
        - transform: A.CLAHE
        - transform: A.ImageCompression
          quality_lower: 50

Thanks!

[Scitator]

Yup, that looks like correct.